"Some people say that we are Nvidia's next billion dollar business," said Kimberly Powell, Vice President of Nvidia Healthcare, in a recent interview. She stated that Nvidia's goal is to provide chips, cloud infrastructure, and other tools for more biotechnology companies.
For Nvidia, healthcare is not its most eye-catching business, but that doesn't mean it doesn't value this area.
If according to the industry classification on NVIDIA's official website, there will be 90 events related to healthcare/life sciences at the upcoming NVIDIA GTC conference on the 18th, ranking first in all industries and surpassing popular fields such as automotive, cloud services, hardware/semiconductors.
Among them, the "The Role of AIGC in Modern Medicine" event hosted by Kimberly Powell brought together renowned fund managers such as Cathie Wood and Microsoft Research Director Peter Lee to explore the key driving forces behind redefining medical services, discovering new drugs, and improving patient efficacy.
As the helmsman of NVIDIA, Huang Renxun has repeatedly referred to digital biology as the "next amazing revolution" in the field of technology. He also stated at a recent conference that the era of computers is over, and human biology is the future.
At the intersection of AIGC and medicine, AI does not write about the ups and downs of life, nor does it depict the scenery of mountains and rivers. It generates the next "miracle medicine" based on each prompt word.
Why is it now?
In most cases, when people talk about Nvidia, they define it as a chip company; The AI craze that began in 2022 has further deepened this impression.
But Huang Renxun does not agree with this. In a 2008 interview, the NVIDIA leader, who always displayed himself in black leather, admitted that,
"On the surface, we do indeed produce chips, but in my eyes, we have never been a chip company, but a company that helps customers solve complex visual computing problems."
"If Nvidia limits itself to a company that produces chips, we will automatically assume that movies have nothing to do with us, games have nothing to do with us, hospitals have nothing to do with us. In fact, these have all become important businesses for us because they have the same problem behind them - complex visual computing problems, none of which belong to competitors, all of which belong to customers."
It can be seen that at that time, Nvidia had already incorporated healthcare into its own territory. Indeed, applying AI to drug discovery is not a new thing. So why, after so many years, tech giants led by Nvidia have started to shout for AI medicine again? Why is it now?
"(At present) is a breakthrough moment." - This is the common answer given by executives from DeepMind and Nvidia. For the first time, the industry has simultaneously integrated three elements: "a large amount of training data, explosive growth of computing resources, and progress in AI algorithms." "This was impossible to achieve five years ago."
AI+Medicine=Next Golden Race?
GPU is the one that has propelled NVIDIA to the "king of computing power", but over the past two years, a significant portion of its venture capital investment has been directed towards drug development - in 2023 alone, NVIDIA invested in eight drug discovery startups.
Figure | NVIDIA 2023 Venture Capital Distribution (Source: S&P Global)
Thanks to the AI boom, Nvidia's market value has rapidly increased. "Since the computer-aided design industry has won the first chip company with a market value of $2 trillion, why can't the computer-aided drug discovery industry create the next trillion dollar drug company?" Kimberly Powell, Vice President of Nvidia Healthcare, explained Nvidia's investment in life sciences.
On the path of AI medicine, Nvidia is not alone - it can be said that technology giants are all interested in AI technology in the field of biomedicine, and the world's most powerful technology giants such as Microsoft and Google are also considering biotechnology as the next frontier of AI.
For example, researchers at Google DeepMind used the AlphaFold model (a breakthrough tool for predicting protein structure) to develop a "molecular" injector that directly injects drugs into cells and is used to study reducing pesticide dependence in agriculture;
Salesforce launched the protein generation model ProGen last year;
Microsoft has also released a similar open-source model, EvoDiff;
Amazon has released a protein folding tool for its AWS machine learning platform SageMaker
Taking DeepMind's AlphaFold research project as an example: Proteins in the human body manage various functions, all of which depend on the three-dimensional shape of proteins. Each protein is composed of a series of amino acids, and the interaction between amino acids and the external environment determines the protein's folding, which determines its final shape.
For biotechnology companies, being able to predict protein shape based on their amino acid sequences is crucial. These companies can use these predictions to design various products such as new drugs, improved crops, and biodegradable plastics.
This is where deep learning comes in handy: training artificial intelligence models on billions of different protein sequences and their underlying structures, allowing these models to discover biological laws without the expensive calculations required for true molecular dynamics simulations. However, fully simulating proteins requires high-intensity computing resources, so some institutions have specifically designed and built supercomputers to handle such problems.
Undoubtedly, AI has enormous potential in the field of biotechnology.
Huang Renxun, who had already turned his attention to medicine 15 years ago, was amazed that the discovery of computer-aided drugs was indeed a miracle. In the field of drug discovery, people can shift from computer-aided drug discovery to computer-aided drug design using the same method as computer-aided chip design. "If we magnify it a billion times, we can simulate biology."
"In the future, life sciences will be highly engineered like traditional industries. When data science, artificial intelligence, and automation are combined, biological societies will improve exponentially, becoming the next golden race track."